Once again, you've written an extremely thought-provoking article. Thank you for putting this together!
I've had T1D for 45 years. I use Control IQ and I (1) love it and (2) don't disagree with your conclusions in the least. I'm meticulous about bolusing for food and high blood sugars. Without that effort, the pump wouldn't be effective at all.
Why do I love it? For me, it's COMPLETELY eliminated overnight lows since the day I switched to CIQ. I used to run high at night out of fear of those overnight lows. No longer! I wake up every morning at 110 mg/dl. This is the one area where closed loop tech is (understandably) excellent. When a diabetic is asleep, the pump can simply release basal insulin, adding more or giving less as needed. Without meals or exercise or stress to worry about when the patient is asleep, it's pretty hard for CIQ to mess up. That alone is worth it for me. But nothing I'm saying contradicts your conclusions.
Thanks again. I look forward to reading each of your posts.
You raise a good point that I originally had in my text, but I removed it later since I was unable to find any third party studies that confirmed the universality of this, and stats on how common it is... I typically strive to avoid anecdotal claims, since many of them are too subjective and unverifiable, and may not represent population-wide experiences.
That said, I'm well aware of that potential benefit, and may go back and insert the idea somewhere in the text sometime if I can find something in the medical literature. (Shouldn't be too hard.)
Thank you for that. The Tslim is the one who’s algorithm has worked best for me in the past, as well as just how the sites work, and whenever I really think about going back to a pump the reasons you decided are absolutely the main ones. I just still struggle with exercise on the pump in fear that it’ll kill me. But thank you for your perspective.
I exercise a lot, and my rule of thumb is to significantly dial back insulin beforehand. How much I reduce and for how long varies, depending on the type and duration of exercise. Everyone's metabolism is different, and your own exertion:insulin ration will also change as your aerobic fitness improves. I have known athletes who don't take any insulin at during exercise, but it took time to get there. (I personally go about 3-6 hours without insulin onboard when I do day-long hikes, while also having to eat. But when I'm done, my insulin requirement slingshots right back up as my metabolism shifts from exertion to rest.)
I would not follow what others do (including me), only to be aware of the kinds of outcomes that can happen. Instead, start with trial and error to get a feeling for it. I always recommend people start simply with 15-20 minute walks directly after meals, and then taper back the insulin if you go hypo for the next time. Doing exercise on a very regular basis really helps in the learning curve.
Naturally, this is a highly complex topic, so there's quite a bit to learn and understand, but this doesn't mean the approach I suggested above won't work. Fear of hypos is the strongest deterrent for T1Ds to exercise, but you don't need to know all the basic science to get it right (but it does help). The best way to get over that is to experimentally taper back insulin as you experiment with building a routine.
Thanks, Dan! I’ve actually been a regular exerciser for two days of free weights, and five or six of walks of 4 to 6 miles, for about the last 40 years. And always at the same time since I keep a very regular schedule. It has luckily just been part of my lifestyle. But I am a morning exerciser and then any later in the day I don’t really count as exercise but I’m aware of it like going on leisure walks. And my weight has been stable so I do have a certain amount of balance. The problem is that I exercise after having my coffee with some sugar-free delight creamer, which raises it enough in the summertime to go the mileage, and usually not have to eat glucose., But in the winter time, my sugar can drop like a rock so I have to feed it more but I check all along the route and have to make stops sometimes at churches and places with seating while I’m on my route. And I rise at 4:15 AM and I’m usually out the door by 515 or 530 so with pumps I just don’t have that much time to turn off things and ultimately I think I’m just going to be an MDI person. It makes me feel safer to carry the insulin in me, rather than on me, as they say. And this winter has been colder than normal, so I think I’m experiencing new methods for preloading with some carbs and which type. And I have a post exercise high, but I think that since I have breakfast after I get back it’s kind of figured in my breakfast ratio somehow since it’s pretty regular. I do sometimes wonder if somehow a person needs some insulin before exercise, as well as carbs, to help fuel the muscles. I love anatomy and physiology as it applies to exercise and hope to continue to figure out what will work. And I’ll know in the future to pay attention as seasons change, and alter what works. Nothing like sitting and freezing on a bench trying to wait for your blood sugar to go up after drinking apple juice or eating glucose tabs and finally getting impatient and calling someone to come get me for them to just rise really high. But I just do not want lows because I think they do trigger more lows and interfere with me trying to be as level as possible. Anyway, I sure do appreciate all of your work to share the data and your experiences, and especially the fact of just kinda listening to my own self as I take in information from other peoples experiences. Sometimes, when reading online, I forget that other people might be taking a whole lot more insulin and may not be doing exercise and have totally different lifestyles. But I sure am getting wise now! Thank you so much for your help! And when I do travel, it always involves hikes in the mountains so your help above with that is perfect, and seems to match what I need to do for that type of exercise. Have a great day!
It's funny. I was a highly competitive tennis player through college and I moved to a pump because it gave me far better control during matches. I lost one of the biggest matches of my life due to a severe hypo event, which persuaded me to switch to a pump (this was back in 1991!). I still work out 5x/week and don't have issues.
I'm not saying this to persuade you or anyone else. My point is that it works for me.
As per the thesis of the article, one can choose to adjust dosing accordingly with MDI as well as pumps. the pump doesn't make it easier--it's more the decision to do it. That, and mastering the personalized insulin-response rate that your body demands. I use an insulin pen and achieve literally the same results, but find the pen far easier to manage than the pump (when I was on it). Everyone's mileage varies.
Humalog didn't exist in 1991. So I was injecting essentially old school Regular insulin. I was the third match on the court in a big tournament and a player in the second match retired a few games in. Meaning I started at least an hour or so earlier than planned. So I took the court as my Regular injection was spiking...and that insulin stayed active far longer than modern insulins. At that time, the ability of a pump's basal rate to give flexibility for things like meals and exercise couldn't be matched by MDI.
I realize things are far, far different now. And it's better for diabetics to have choices. But the pump was such a life changer for in in the 90s (particularly starting college, sleeping late, going out late, etc) I can't imagine switching back. The pump freed me from have a strict schedule.
The thing about T1D is that it's so essential to have the routine that works for you, and half the struggle is just the mental agility--the ability to focus on it in whatever way that works. Whenever anyone gets a routine that "works" (achieve healthy glycemic control), there's never a reason to change it. One of my pet peeves is seeing clinicians recommend to ideally controlled T1Ds that they change something, like new insulins or technologies.
The reason why something works is not because of the thing--it's because the person found a way to use it to achieve healthy outcomes.
“Another way to test that is by measuring A1c levels before and after introducing people to pumps and measuring their outcomes afterwards (while including subjects from diverse social, racial and economic subpopulations). This metaanalysis (literature review) of dozens of RTCs had these aims in mind, and also included”
Please, sir, what are RTCs?
Policing over. Again, thank you for an interesting article. I always learn something from your posts. I'm a Metronic user and I got a new pump over a year ago. They gave me a box of sensors and a chance to try out this new 780G system. Well, I have a problem with rising blood sugar after I get out of bed. I like to eat right away, so first thing I do is check my numbers and enter carbs for breakfast. However, auto mode thought that my rising blood sugars were from eating, but I knew they weren't. So I had to exit auto mode, bolus in manual mode, and then go back to auto mode a few hours later for the rest of the day. I learned from a forum that I could "phantom bolus" to correct, but it just seemed a hassle. It was way easier to set my Basals where they needed to be. Plus, I would have to pay out of pocket for the sensors, when the Libre 2 was covered by the government.
I missed my old pump, because the new one was specifically designed to be used in auto mode and manual mode requires seemingly endless button pressing. Unfortunately, I forgot to take that old pump off when I went swimming in the sea. "Critical Error!" ;-) So, back to my "back up", newer pump.
My 2 months on the Medtronic "closed loop" system was very frustrating and my time in range numbers were worse. I can see where it would be advantageous for someone who was hypo unaware, especially at night, but it just wasn't working for me.
thanks for the spell/grammar checking! It's so much nicer when one has an editor, but alas, I rely on my audience. :-)
RTC was defined earlier in the article, but I redefined it again where you indicated (randomized control trials).
Your experience illustrates the value of knowing how to manage T1D and intervening when appropriate. Many think they won't need to do that, leading them to stop paying attention, thereby leading to worse outcomes.
Thanks so much for combing through the studies’ detail and summarizing smartly for us. I am personally so grateful since I have for 2 years tried hard to stick with a pump and yet conclusively proved to myself with a full 90-day MDI experience (post pumps) that MDI is more liberating and educational and effective. I had an amazing 5.6 A1C and proved my instincts are right. Plus, I do keep learning, which helps alleviate fears of lows and navigate the exercise-related challenges in different scenarios like seasons and temperature and such. I feel like pumping was sort of favored by my health practitioners, though not any longer as they now let me experiment with very fast-acting insulins, dosing regimens and really support anything I want to try (I am deeply motivated to stay healthy at age 60 and to find balance, which I am now working toward by trying to let go all my daily record-keeping and ritual of data-study every morning to make basal dosing decisions and such). This article and, indeed, all that I have read here, feels so supportive of my innate approach to diabetes management. I have felt asea on some pump-oriented forums, and even “bold with insulin” approaches that have led to some very real scares. Thank-you so much for the wisdom, authenticity and help here. I have also been struggling with decisions on Freestyle Libre 3 (always accurate to 5 points compared to meter - though should be a lag, I suppose) versus Dexcom g6 (needing multiple calibrations this whole past year, but love the data, Clarity App and realize the frequent reporting of L3 can cause rash wrong decisions versus every 5 minutes). Still trying to see which one will be better and what I would give up. G6 constantly runs 40 points low during and after exercise in cold weather, but can I learn to just react with glucose at a different, so-called wrong number and which will give better peace of mind? Crazy, but I think g6 could lead to smoother life, though hate finger-sticking. So all that wordiness to say, I am beyond grateful how all of this that you give us feeds me so personally as I try to learn. I am only 3 years into this drastic change of life and how I can proceed and become free within its constraints and let go fear. I am lucky to have always prized health and had my exercise, water and meal-times already set as stable. I don’t have time to err too much and have needed this information and perspective to be shared with me. Sorry, this was so long winded, but just want to share my appreciation and how personal this is. Thank you so much.
the big news item in your post is that you're only 3yrs into T1D, which means that your body still has a lot of transitioning to do. It really takes a lot of time for all the other hormonal imbalances to catch up to the fact that the beta cells aren't producing insulin--it doesn't all just happen at once. As glycemic control goes beyond "normal range," those counter-regulatory systems stress and eventually fatigue and are less effective. Alpha-cell response is a good example (which produces glucagon when you go low). If you can keep your A1c low while also minimizing lows (<70), you can keep many of those other counter-regulatory responses from diminishing.
Your experience with L3 vs G6 is interesting--I've found the opposite: the L3 reports numbers ~20 lower than the G6. The greatest advantage to dexcom is calibration, which was a primary frustration with the Lw and L3. I also didn't like that the Abbott products had jumpy readings. I'm sure you've read my article comparing the G6 and G7, and there, I talk about the fact that glucose concentrations in the body are not evenly distributed, and it's particularly difficult to get a true "systemic" glucose level from a single reading (including from BGMs). The greatest value of the G6 is that its algorithm takes this into account, and adjusts its data in a manner that gives a much more precise measure of systemic glucose levels--and it's trajectory--which can be observed in shorter time windows, allowing you to react more quickly with adjustments as necessary.
Thanks, Dan! And the funny thing is that overnight the libre three always from 10 P.m. until I get up about 415 is at least 20 points below the Dexcom. After I take this libre three out when it expires in about 10 days, I am just going to stick with the Dexcom G6 because it is smoother and doesn’t cause stress when I look at it like with changes in the Libre. Plus their customer service is much nicer about replacing sensors that don’t work or even those that come out early. I sure hope they’ll keep the G6 around. And if you have any anecdotal or other knowledge, could you comment about presoaking and when to calibrate the first time on a new one, assuming it’s necessary when you do the fingerstick? I am starting to pre-soak for 24 hours and last time I ended up calibrating at about the two hour mark because it was so far off like 40 points, but somewhere I read not to calibrate for the first 24 hours. What I’m finding is that that calibration worked well and then about day four or five I do another and I’m on day 10 right now I’m going to change it after lunch and it’s working great even for the last day. So this time I thought I might not even wait two hours to calibrate and just do it after one hour if it’s really Squirrley looking. Final comment on the Libre three is that it really does have compression lows, no matter where I put it on my arm, overnight every single time in my experience. I think it could be from the size of it. And the presoaking is a weird one for me because I do fear I’ll have a bleeder and I wouldn’t want to leave that open and lately I’ve used a product called press and seal, which is kind of like cling wrap, but stays on skin very well and I’ve put it over the open presoaking sensor just thinking I don’t want germs in there, but in the summer when I did it, it made condensation and for today I actually just took the cover off while doing weight training and I’m leaving it off till I change the sensor after lunch. I sort of hate adding other variables to all of this, but somehow it seems like soaking, could actually help the tissue get normalized for better readings. Thanks in advance if you have time to respond on the these two questions. And thanks for the above. I feel like I’m probably the most motivated person on the planet to manage my BG and do study my clarity data daily to inform Tresiba amount and pick up on trends. So I really appreciate what you said above about being so new to this, and just having assumed that the honeymoon’s over, so that’s that, while there are actually many more hormones and that I have a chance to keep the counter-regulatory responses from diminishing further. It always feels great to think I can really make a good impact on all of this. Thanks again.
Hi Dan, this is an interesting article, but I feel as though you're conflating two separate issues. Pump therapy and automation, which I think are two different (but linked) items.
The majority of the studies showing greater issues with pump therapy that you've referred to talk about traditional pumps rather than Automated Insulin Delivery devices. If users consider traditional CSII to be automated devices rather than devices that replace MDI with a more flexible delivery system that still requires equivalent levels of user interaction (meals, exercise and illness being the key ones), then who is at fault? The user or those who provide the alternative delivery system to them without appropriate introduction and education?
Automation without knowledge of what to do without it, on the other hand, does present potential problems. Even then, nothing that's on the market right now should be considered as anything other than hybrid closed loop, requiring intervention for both meals and exercise, and if users aren't using it that way, then again we come back to the question of why not? Is it down to the user, the technology or the environment in which it is deployed?
While the person manages the diabetes, if technology frees up headspace requiring less user oversight and results in the same numbers, I don't see that as an issue, even if one is capable of taking care of themselves to a high level without using it.
As the article points out, there are major psychological factors to consider: "Freeing up headspace," as you call it, actually does more harm than good. When you're not paying close attention, it's easy to miss signals or take actions--or tolerate being bothered--then your performance degrades. Hence, The Hawthorne Effect: If you're forced to pay attention, as all these cited studies show, then people's self-management is better.
Obviously, this isn't universally true--there are those who do just as well with pumps (including semi-automated) as MDI--but studies show these people generally don't do better... mostly just the same.
There are also other factors, such as cost and technical errors (the so-called "futz factor" associated with complicated technology) that all the researchers cited in this article attribute to being a lower-quality experience.
All that said, there's no discounting the phenomenon known as the Christmas Tree Effect, which the article concedes is a major factor in how people think about insulin pumps. They're kinda cool and high tech, which has a big effect on how much they (think they) like it.
Some of this reminds me of findings about autopilots in the airline industry showing that pilots not having enough to do leads to inattention that in turn causes problems when intervention by the pilot is necessary. But that leads to something that's kind of a hobby-horse for me (T1 since 1983) around AID systems and that gets touched on here. Namely, is the ideal to have a black-box, set-and-forget system that does everything for you? That seems to be the assumption in this article, but my experience with these systems, having been on both Medtronic and Tandem, is that there are two design philosophies in play. The Tandem seems to be more a kind of smart assistant that offers options and can do corrections when things get off balance but doesn't want to assume complete control, where the Medtronic approach was very much more that of a HAL9000 ("I'm sorry Dave. I can't let you do that.") I think the Tandem approach is much more in line with inherent limitations of the whole problem.
Other stuff:
> CGM vs Pump. I've said many times that if I had to choose between the two, I'd give up the pump before I gave up the CGM hands down. So yeah. But I still think there are real quality of life considerations. Such as....
What pumps are really for: solving the basal conundrum.
I was on R/NPH for 20 yrs, then Basal/Bolus MDI for 10 before starting with a pump in 2004, pre-CGM. Until then I was NEVER able to get ahead of Dawn Phenomenon, and struggled hugely to accommodate exercise, unanticipated overnight excursions and the like. Even with a "dumb" pump, being able to program steps in your basal delivery to accommodate more-or-less predictable diurnal changes in your basal BG output is a huge lifestyle improvement as well as management. You can't turn Lantus OFF when you want to go for a bike ride or have to run to catch a bus. To me this flexibility is 80% of the justification, leaving aside Christmas tree bells and whistles. I was surprised not to see this given more emphasis.
Outcomes vs quality-of-life
I was pleased to see q-o-l treated as a significant consideration here, as it is so often left out in favor of "outcomes" that are more reducible to numbers. The basal control issue is the most significant for me in terms of this. I grant that there are many more options now than good old Lantus, but essentially they all have the same problem: the "resolution" can only vaguely conform to diurnal cycles, which are largely predictable and thus amenable to a programmable delivery system, even a "dumb" one, and being able to adjust in light of unforeseen events is a huge benefit to leading something resembling a normal life. Like I say, I was on R/NPH, which for good reason I used to refer to as the "East now or DIE!" regimen, so I'm very sensitive on this issue. If the data says my A1C isn't all that much better I'd still respond that A1C isn't the only measure I care about.
A last thing I don't see here that I think is worth mentioning: insulin speed. I've gone from Regular to Novolog/Humalog to Fiasp. Fiasp is noticeably faster for me, but there's still a frustratingly long lag time, especially when it comes to correcting a post-prandial that's gotten out of control. I know people who speak highly of Afrezza but inhalable insulin can have other problems. Mainly I'm thinking about the flaws with AID and pumps in general, where the issue for me hasn't really been about the programming or CGM accuracy as just how slowly **any** of these systems can respond to BG changes due to the effect curve of the insulin itself as well as the inherent inefficiancy of subcutaneous injection vs a functional pancreas attached to your bloodstream. To me this is the limiting factor in the way of a true "black-box" device and why I prefer a system that doesn't overpromise what it can actually achieve.
Anyway, thanks again--lots to think about and links to follow up here. This represents a lot of work pulling sources together and I think it's a real contribution to our understanding.
Thanks for this detailed comment. There's a lot there, but I didn't approach many of these issues for a variety of reasons. Primarily (if not solely), I wanted to remain focused on the epidemiological studies across populations, especially when conducted by independent researchers using quality study methods (randomized control trials).
Yes, there are a lot of "management techniques" outside of that domain, and many of your examples are exactly that, which speaks to my commentary at the end of the article: If you know what you're doing, you can translate that into whatever insulin delivery method you use. And the more informed you are and more in tune with your body's physiology, the more successful you'll be at it.
Your comment about basal insulin needs is important, and I was debating on whether (and to what degree) I would incorporate this into the article, but I ultimately decided against it because it is more related to management techniques that isn't part of the population-wide studies. I will be writing a separate article about insulin absorption and how it is a huge factor in glycemic management, which spans not just insulin types, but delivery method, location, quantity, metabolization, and other factors.
As it pertains to pumps, your argument that they are better than most slow-acting insulins is probably the most common and popular explanation people give about why they love pumps. But again, it's not that simple. Most people's basal needs are not what they think they are, largely because one's basal needs are not nearly as "flat" throughout the day (and night) as they think they are. If you're not the type of fine-tune your management, you're more likely to just follow your endo's recommendation and keep it largely flat. This will mean that people are often getting more insulin (in the form of basal) at certain times of day than they need, so they end up eating more carbs than they would otherwise do, which leads to weight gain and a downward spiral of insulin resistance. (See my previous articles "HbA1c Tests and T1D: The Good, The Bad and the Ugly" and "Why Controlling Glucose is so Tricky" for more on this.) Night basal needs are highly volatile as well--the more exercise you do, the lower your nighttime basal needs. Also, as people age (>50), your basal needs at night really start to drop.
At the end of the day, the complexity of basal is such that the more important factor is knowing your own personal variability. If you watch your CGM and your carb::insulin ratios during the day and night, you can figure that out. Granted, it requires time and attention, and yes, it's the hardest part of the whole thing. But if you can figure that out, you can manage it just as easily MDI than on a pump. And if you can't figure out, then your outcome won't be any different on a pump than MDI.
And this is what brings us full circle: Those who are in best control are those who took the initiative to learn and focus on the details of their T1D. Due to the various psychological factors discussed in the article, the attraction of pumps also has the detrimental effect of keeping people from focusing on those nuanes. The Hawthorne Effect finds that people who engage in MDI tend to be more attentive to detail than pump users. This is what the studies show, but of course, individuals have the power to overcome these factors.
Once again, you've written an extremely thought-provoking article. Thank you for putting this together!
I've had T1D for 45 years. I use Control IQ and I (1) love it and (2) don't disagree with your conclusions in the least. I'm meticulous about bolusing for food and high blood sugars. Without that effort, the pump wouldn't be effective at all.
Why do I love it? For me, it's COMPLETELY eliminated overnight lows since the day I switched to CIQ. I used to run high at night out of fear of those overnight lows. No longer! I wake up every morning at 110 mg/dl. This is the one area where closed loop tech is (understandably) excellent. When a diabetic is asleep, the pump can simply release basal insulin, adding more or giving less as needed. Without meals or exercise or stress to worry about when the patient is asleep, it's pretty hard for CIQ to mess up. That alone is worth it for me. But nothing I'm saying contradicts your conclusions.
Thanks again. I look forward to reading each of your posts.
"it's COMPLETELY eliminated overnight lows..."
You raise a good point that I originally had in my text, but I removed it later since I was unable to find any third party studies that confirmed the universality of this, and stats on how common it is... I typically strive to avoid anecdotal claims, since many of them are too subjective and unverifiable, and may not represent population-wide experiences.
That said, I'm well aware of that potential benefit, and may go back and insert the idea somewhere in the text sometime if I can find something in the medical literature. (Shouldn't be too hard.)
Thanks for bringing it (back) to my attention!
Thank you for that. The Tslim is the one who’s algorithm has worked best for me in the past, as well as just how the sites work, and whenever I really think about going back to a pump the reasons you decided are absolutely the main ones. I just still struggle with exercise on the pump in fear that it’ll kill me. But thank you for your perspective.
I exercise a lot, and my rule of thumb is to significantly dial back insulin beforehand. How much I reduce and for how long varies, depending on the type and duration of exercise. Everyone's metabolism is different, and your own exertion:insulin ration will also change as your aerobic fitness improves. I have known athletes who don't take any insulin at during exercise, but it took time to get there. (I personally go about 3-6 hours without insulin onboard when I do day-long hikes, while also having to eat. But when I'm done, my insulin requirement slingshots right back up as my metabolism shifts from exertion to rest.)
I would not follow what others do (including me), only to be aware of the kinds of outcomes that can happen. Instead, start with trial and error to get a feeling for it. I always recommend people start simply with 15-20 minute walks directly after meals, and then taper back the insulin if you go hypo for the next time. Doing exercise on a very regular basis really helps in the learning curve.
Naturally, this is a highly complex topic, so there's quite a bit to learn and understand, but this doesn't mean the approach I suggested above won't work. Fear of hypos is the strongest deterrent for T1Ds to exercise, but you don't need to know all the basic science to get it right (but it does help). The best way to get over that is to experimentally taper back insulin as you experiment with building a routine.
Thanks, Dan! I’ve actually been a regular exerciser for two days of free weights, and five or six of walks of 4 to 6 miles, for about the last 40 years. And always at the same time since I keep a very regular schedule. It has luckily just been part of my lifestyle. But I am a morning exerciser and then any later in the day I don’t really count as exercise but I’m aware of it like going on leisure walks. And my weight has been stable so I do have a certain amount of balance. The problem is that I exercise after having my coffee with some sugar-free delight creamer, which raises it enough in the summertime to go the mileage, and usually not have to eat glucose., But in the winter time, my sugar can drop like a rock so I have to feed it more but I check all along the route and have to make stops sometimes at churches and places with seating while I’m on my route. And I rise at 4:15 AM and I’m usually out the door by 515 or 530 so with pumps I just don’t have that much time to turn off things and ultimately I think I’m just going to be an MDI person. It makes me feel safer to carry the insulin in me, rather than on me, as they say. And this winter has been colder than normal, so I think I’m experiencing new methods for preloading with some carbs and which type. And I have a post exercise high, but I think that since I have breakfast after I get back it’s kind of figured in my breakfast ratio somehow since it’s pretty regular. I do sometimes wonder if somehow a person needs some insulin before exercise, as well as carbs, to help fuel the muscles. I love anatomy and physiology as it applies to exercise and hope to continue to figure out what will work. And I’ll know in the future to pay attention as seasons change, and alter what works. Nothing like sitting and freezing on a bench trying to wait for your blood sugar to go up after drinking apple juice or eating glucose tabs and finally getting impatient and calling someone to come get me for them to just rise really high. But I just do not want lows because I think they do trigger more lows and interfere with me trying to be as level as possible. Anyway, I sure do appreciate all of your work to share the data and your experiences, and especially the fact of just kinda listening to my own self as I take in information from other peoples experiences. Sometimes, when reading online, I forget that other people might be taking a whole lot more insulin and may not be doing exercise and have totally different lifestyles. But I sure am getting wise now! Thank you so much for your help! And when I do travel, it always involves hikes in the mountains so your help above with that is perfect, and seems to match what I need to do for that type of exercise. Have a great day!
It's funny. I was a highly competitive tennis player through college and I moved to a pump because it gave me far better control during matches. I lost one of the biggest matches of my life due to a severe hypo event, which persuaded me to switch to a pump (this was back in 1991!). I still work out 5x/week and don't have issues.
I'm not saying this to persuade you or anyone else. My point is that it works for me.
Same here--but with racquetball!
As per the thesis of the article, one can choose to adjust dosing accordingly with MDI as well as pumps. the pump doesn't make it easier--it's more the decision to do it. That, and mastering the personalized insulin-response rate that your body demands. I use an insulin pen and achieve literally the same results, but find the pen far easier to manage than the pump (when I was on it). Everyone's mileage varies.
Humalog didn't exist in 1991. So I was injecting essentially old school Regular insulin. I was the third match on the court in a big tournament and a player in the second match retired a few games in. Meaning I started at least an hour or so earlier than planned. So I took the court as my Regular injection was spiking...and that insulin stayed active far longer than modern insulins. At that time, the ability of a pump's basal rate to give flexibility for things like meals and exercise couldn't be matched by MDI.
I realize things are far, far different now. And it's better for diabetics to have choices. But the pump was such a life changer for in in the 90s (particularly starting college, sleeping late, going out late, etc) I can't imagine switching back. The pump freed me from have a strict schedule.
The thing about T1D is that it's so essential to have the routine that works for you, and half the struggle is just the mental agility--the ability to focus on it in whatever way that works. Whenever anyone gets a routine that "works" (achieve healthy glycemic control), there's never a reason to change it. One of my pet peeves is seeing clinicians recommend to ideally controlled T1Ds that they change something, like new insulins or technologies.
The reason why something works is not because of the thing--it's because the person found a way to use it to achieve healthy outcomes.
Proofreading Police! 😉
“Technically, such a system a highly“
You’re missing a verb, perhaps “is”?
“Another way to test that is by measuring A1c levels before and after introducing people to pumps and measuring their outcomes afterwards (while including subjects from diverse social, racial and economic subpopulations). This metaanalysis (literature review) of dozens of RTCs had these aims in mind, and also included”
Please, sir, what are RTCs?
Policing over. Again, thank you for an interesting article. I always learn something from your posts. I'm a Metronic user and I got a new pump over a year ago. They gave me a box of sensors and a chance to try out this new 780G system. Well, I have a problem with rising blood sugar after I get out of bed. I like to eat right away, so first thing I do is check my numbers and enter carbs for breakfast. However, auto mode thought that my rising blood sugars were from eating, but I knew they weren't. So I had to exit auto mode, bolus in manual mode, and then go back to auto mode a few hours later for the rest of the day. I learned from a forum that I could "phantom bolus" to correct, but it just seemed a hassle. It was way easier to set my Basals where they needed to be. Plus, I would have to pay out of pocket for the sensors, when the Libre 2 was covered by the government.
I missed my old pump, because the new one was specifically designed to be used in auto mode and manual mode requires seemingly endless button pressing. Unfortunately, I forgot to take that old pump off when I went swimming in the sea. "Critical Error!" ;-) So, back to my "back up", newer pump.
My 2 months on the Medtronic "closed loop" system was very frustrating and my time in range numbers were worse. I can see where it would be advantageous for someone who was hypo unaware, especially at night, but it just wasn't working for me.
I look forward to your next article.
thanks for the spell/grammar checking! It's so much nicer when one has an editor, but alas, I rely on my audience. :-)
RTC was defined earlier in the article, but I redefined it again where you indicated (randomized control trials).
Your experience illustrates the value of knowing how to manage T1D and intervening when appropriate. Many think they won't need to do that, leading them to stop paying attention, thereby leading to worse outcomes.
Thanks so much for combing through the studies’ detail and summarizing smartly for us. I am personally so grateful since I have for 2 years tried hard to stick with a pump and yet conclusively proved to myself with a full 90-day MDI experience (post pumps) that MDI is more liberating and educational and effective. I had an amazing 5.6 A1C and proved my instincts are right. Plus, I do keep learning, which helps alleviate fears of lows and navigate the exercise-related challenges in different scenarios like seasons and temperature and such. I feel like pumping was sort of favored by my health practitioners, though not any longer as they now let me experiment with very fast-acting insulins, dosing regimens and really support anything I want to try (I am deeply motivated to stay healthy at age 60 and to find balance, which I am now working toward by trying to let go all my daily record-keeping and ritual of data-study every morning to make basal dosing decisions and such). This article and, indeed, all that I have read here, feels so supportive of my innate approach to diabetes management. I have felt asea on some pump-oriented forums, and even “bold with insulin” approaches that have led to some very real scares. Thank-you so much for the wisdom, authenticity and help here. I have also been struggling with decisions on Freestyle Libre 3 (always accurate to 5 points compared to meter - though should be a lag, I suppose) versus Dexcom g6 (needing multiple calibrations this whole past year, but love the data, Clarity App and realize the frequent reporting of L3 can cause rash wrong decisions versus every 5 minutes). Still trying to see which one will be better and what I would give up. G6 constantly runs 40 points low during and after exercise in cold weather, but can I learn to just react with glucose at a different, so-called wrong number and which will give better peace of mind? Crazy, but I think g6 could lead to smoother life, though hate finger-sticking. So all that wordiness to say, I am beyond grateful how all of this that you give us feeds me so personally as I try to learn. I am only 3 years into this drastic change of life and how I can proceed and become free within its constraints and let go fear. I am lucky to have always prized health and had my exercise, water and meal-times already set as stable. I don’t have time to err too much and have needed this information and perspective to be shared with me. Sorry, this was so long winded, but just want to share my appreciation and how personal this is. Thank you so much.
the big news item in your post is that you're only 3yrs into T1D, which means that your body still has a lot of transitioning to do. It really takes a lot of time for all the other hormonal imbalances to catch up to the fact that the beta cells aren't producing insulin--it doesn't all just happen at once. As glycemic control goes beyond "normal range," those counter-regulatory systems stress and eventually fatigue and are less effective. Alpha-cell response is a good example (which produces glucagon when you go low). If you can keep your A1c low while also minimizing lows (<70), you can keep many of those other counter-regulatory responses from diminishing.
Your experience with L3 vs G6 is interesting--I've found the opposite: the L3 reports numbers ~20 lower than the G6. The greatest advantage to dexcom is calibration, which was a primary frustration with the Lw and L3. I also didn't like that the Abbott products had jumpy readings. I'm sure you've read my article comparing the G6 and G7, and there, I talk about the fact that glucose concentrations in the body are not evenly distributed, and it's particularly difficult to get a true "systemic" glucose level from a single reading (including from BGMs). The greatest value of the G6 is that its algorithm takes this into account, and adjusts its data in a manner that gives a much more precise measure of systemic glucose levels--and it's trajectory--which can be observed in shorter time windows, allowing you to react more quickly with adjustments as necessary.
Thanks, Dan! And the funny thing is that overnight the libre three always from 10 P.m. until I get up about 415 is at least 20 points below the Dexcom. After I take this libre three out when it expires in about 10 days, I am just going to stick with the Dexcom G6 because it is smoother and doesn’t cause stress when I look at it like with changes in the Libre. Plus their customer service is much nicer about replacing sensors that don’t work or even those that come out early. I sure hope they’ll keep the G6 around. And if you have any anecdotal or other knowledge, could you comment about presoaking and when to calibrate the first time on a new one, assuming it’s necessary when you do the fingerstick? I am starting to pre-soak for 24 hours and last time I ended up calibrating at about the two hour mark because it was so far off like 40 points, but somewhere I read not to calibrate for the first 24 hours. What I’m finding is that that calibration worked well and then about day four or five I do another and I’m on day 10 right now I’m going to change it after lunch and it’s working great even for the last day. So this time I thought I might not even wait two hours to calibrate and just do it after one hour if it’s really Squirrley looking. Final comment on the Libre three is that it really does have compression lows, no matter where I put it on my arm, overnight every single time in my experience. I think it could be from the size of it. And the presoaking is a weird one for me because I do fear I’ll have a bleeder and I wouldn’t want to leave that open and lately I’ve used a product called press and seal, which is kind of like cling wrap, but stays on skin very well and I’ve put it over the open presoaking sensor just thinking I don’t want germs in there, but in the summer when I did it, it made condensation and for today I actually just took the cover off while doing weight training and I’m leaving it off till I change the sensor after lunch. I sort of hate adding other variables to all of this, but somehow it seems like soaking, could actually help the tissue get normalized for better readings. Thanks in advance if you have time to respond on the these two questions. And thanks for the above. I feel like I’m probably the most motivated person on the planet to manage my BG and do study my clarity data daily to inform Tresiba amount and pick up on trends. So I really appreciate what you said above about being so new to this, and just having assumed that the honeymoon’s over, so that’s that, while there are actually many more hormones and that I have a chance to keep the counter-regulatory responses from diminishing further. It always feels great to think I can really make a good impact on all of this. Thanks again.
Hi Dan, this is an interesting article, but I feel as though you're conflating two separate issues. Pump therapy and automation, which I think are two different (but linked) items.
The majority of the studies showing greater issues with pump therapy that you've referred to talk about traditional pumps rather than Automated Insulin Delivery devices. If users consider traditional CSII to be automated devices rather than devices that replace MDI with a more flexible delivery system that still requires equivalent levels of user interaction (meals, exercise and illness being the key ones), then who is at fault? The user or those who provide the alternative delivery system to them without appropriate introduction and education?
Automation without knowledge of what to do without it, on the other hand, does present potential problems. Even then, nothing that's on the market right now should be considered as anything other than hybrid closed loop, requiring intervention for both meals and exercise, and if users aren't using it that way, then again we come back to the question of why not? Is it down to the user, the technology or the environment in which it is deployed?
While the person manages the diabetes, if technology frees up headspace requiring less user oversight and results in the same numbers, I don't see that as an issue, even if one is capable of taking care of themselves to a high level without using it.
As the article points out, there are major psychological factors to consider: "Freeing up headspace," as you call it, actually does more harm than good. When you're not paying close attention, it's easy to miss signals or take actions--or tolerate being bothered--then your performance degrades. Hence, The Hawthorne Effect: If you're forced to pay attention, as all these cited studies show, then people's self-management is better.
Obviously, this isn't universally true--there are those who do just as well with pumps (including semi-automated) as MDI--but studies show these people generally don't do better... mostly just the same.
There are also other factors, such as cost and technical errors (the so-called "futz factor" associated with complicated technology) that all the researchers cited in this article attribute to being a lower-quality experience.
All that said, there's no discounting the phenomenon known as the Christmas Tree Effect, which the article concedes is a major factor in how people think about insulin pumps. They're kinda cool and high tech, which has a big effect on how much they (think they) like it.
Some of this reminds me of findings about autopilots in the airline industry showing that pilots not having enough to do leads to inattention that in turn causes problems when intervention by the pilot is necessary. But that leads to something that's kind of a hobby-horse for me (T1 since 1983) around AID systems and that gets touched on here. Namely, is the ideal to have a black-box, set-and-forget system that does everything for you? That seems to be the assumption in this article, but my experience with these systems, having been on both Medtronic and Tandem, is that there are two design philosophies in play. The Tandem seems to be more a kind of smart assistant that offers options and can do corrections when things get off balance but doesn't want to assume complete control, where the Medtronic approach was very much more that of a HAL9000 ("I'm sorry Dave. I can't let you do that.") I think the Tandem approach is much more in line with inherent limitations of the whole problem.
Other stuff:
> CGM vs Pump. I've said many times that if I had to choose between the two, I'd give up the pump before I gave up the CGM hands down. So yeah. But I still think there are real quality of life considerations. Such as....
What pumps are really for: solving the basal conundrum.
I was on R/NPH for 20 yrs, then Basal/Bolus MDI for 10 before starting with a pump in 2004, pre-CGM. Until then I was NEVER able to get ahead of Dawn Phenomenon, and struggled hugely to accommodate exercise, unanticipated overnight excursions and the like. Even with a "dumb" pump, being able to program steps in your basal delivery to accommodate more-or-less predictable diurnal changes in your basal BG output is a huge lifestyle improvement as well as management. You can't turn Lantus OFF when you want to go for a bike ride or have to run to catch a bus. To me this flexibility is 80% of the justification, leaving aside Christmas tree bells and whistles. I was surprised not to see this given more emphasis.
Outcomes vs quality-of-life
I was pleased to see q-o-l treated as a significant consideration here, as it is so often left out in favor of "outcomes" that are more reducible to numbers. The basal control issue is the most significant for me in terms of this. I grant that there are many more options now than good old Lantus, but essentially they all have the same problem: the "resolution" can only vaguely conform to diurnal cycles, which are largely predictable and thus amenable to a programmable delivery system, even a "dumb" one, and being able to adjust in light of unforeseen events is a huge benefit to leading something resembling a normal life. Like I say, I was on R/NPH, which for good reason I used to refer to as the "East now or DIE!" regimen, so I'm very sensitive on this issue. If the data says my A1C isn't all that much better I'd still respond that A1C isn't the only measure I care about.
A last thing I don't see here that I think is worth mentioning: insulin speed. I've gone from Regular to Novolog/Humalog to Fiasp. Fiasp is noticeably faster for me, but there's still a frustratingly long lag time, especially when it comes to correcting a post-prandial that's gotten out of control. I know people who speak highly of Afrezza but inhalable insulin can have other problems. Mainly I'm thinking about the flaws with AID and pumps in general, where the issue for me hasn't really been about the programming or CGM accuracy as just how slowly **any** of these systems can respond to BG changes due to the effect curve of the insulin itself as well as the inherent inefficiancy of subcutaneous injection vs a functional pancreas attached to your bloodstream. To me this is the limiting factor in the way of a true "black-box" device and why I prefer a system that doesn't overpromise what it can actually achieve.
Anyway, thanks again--lots to think about and links to follow up here. This represents a lot of work pulling sources together and I think it's a real contribution to our understanding.
Thanks for this detailed comment. There's a lot there, but I didn't approach many of these issues for a variety of reasons. Primarily (if not solely), I wanted to remain focused on the epidemiological studies across populations, especially when conducted by independent researchers using quality study methods (randomized control trials).
Yes, there are a lot of "management techniques" outside of that domain, and many of your examples are exactly that, which speaks to my commentary at the end of the article: If you know what you're doing, you can translate that into whatever insulin delivery method you use. And the more informed you are and more in tune with your body's physiology, the more successful you'll be at it.
Your comment about basal insulin needs is important, and I was debating on whether (and to what degree) I would incorporate this into the article, but I ultimately decided against it because it is more related to management techniques that isn't part of the population-wide studies. I will be writing a separate article about insulin absorption and how it is a huge factor in glycemic management, which spans not just insulin types, but delivery method, location, quantity, metabolization, and other factors.
As it pertains to pumps, your argument that they are better than most slow-acting insulins is probably the most common and popular explanation people give about why they love pumps. But again, it's not that simple. Most people's basal needs are not what they think they are, largely because one's basal needs are not nearly as "flat" throughout the day (and night) as they think they are. If you're not the type of fine-tune your management, you're more likely to just follow your endo's recommendation and keep it largely flat. This will mean that people are often getting more insulin (in the form of basal) at certain times of day than they need, so they end up eating more carbs than they would otherwise do, which leads to weight gain and a downward spiral of insulin resistance. (See my previous articles "HbA1c Tests and T1D: The Good, The Bad and the Ugly" and "Why Controlling Glucose is so Tricky" for more on this.) Night basal needs are highly volatile as well--the more exercise you do, the lower your nighttime basal needs. Also, as people age (>50), your basal needs at night really start to drop.
At the end of the day, the complexity of basal is such that the more important factor is knowing your own personal variability. If you watch your CGM and your carb::insulin ratios during the day and night, you can figure that out. Granted, it requires time and attention, and yes, it's the hardest part of the whole thing. But if you can figure that out, you can manage it just as easily MDI than on a pump. And if you can't figure out, then your outcome won't be any different on a pump than MDI.
And this is what brings us full circle: Those who are in best control are those who took the initiative to learn and focus on the details of their T1D. Due to the various psychological factors discussed in the article, the attraction of pumps also has the detrimental effect of keeping people from focusing on those nuanes. The Hawthorne Effect finds that people who engage in MDI tend to be more attentive to detail than pump users. This is what the studies show, but of course, individuals have the power to overcome these factors.